Article Text

Download PDFPDF

Factors predicting mortality in victims of blunt trauma brain injury in emergency department settings
  1. K-Y Hsiao1,4,
  2. C-T Hsiao1,4,
  3. H-H Weng2,4,6,
  4. K-H Chen3,4,
  5. L-J Lin1,4,
  6. Y-M Huang4,5
  1. 1
    Department of Emergency Medicine, Chang Gung Memorial Hospital, Chiayi, Taiwan, ROC
  2. 2
    Department of Diagnostic Radiology, Chang Gung Memorial Hospital, Chiayi, Taiwan, ROC
  3. 3
    Department of Physical Medicine and Rehabilitation, Chang Gung Memorial Hospital, Chiayi, Taiwan, ROC
  4. 4
    College of Medicine, Chang Gung University, Taoyuan, Taiwan, ROC
  5. 5
    Department of Nursing, Chang Gung Memorial Hospital, Chiayi, Taiwan, ROC
  6. 6
    Graduate Institute of Occupational Safety and Health, College of Health Sciences, Kaohsiung Medical University, Kaohsiung, Taiwan, ROC
  1. Mr Y-M Huang, Department of Emergency Medicine, Chiayi Chang Gung Memorial Hospital, No 6, West Section, Chia-Pu Road, Puzih City, Chiayi County 613, Taiwan, ROC; richard.smith{at}


Objective: The aim of the study was to identify risk factors for mortality in patients brought to the emergency department (ED) after blunt traumatic brain injury (TBI).

Methods: The medical records of such patients who visited the ED from June 2004 to May 2005 were retrospectively reviewed. Data (age, gender, initial Glasgow coma scale (GCS) scores, initial vital signs, brain computed tomography scan findings and cause of trauma) were collected from the records of 204 TBI patients, who were treated at the ED and needed intensive care. Among these patients, 48 died in the intensive care unit (ICU) of the hospital. Logistic regression was used to assess factors affecting mortality after trauma.

Results: Age (odds ratio (OR) 1.04; 95% CI 1.01 to ∼1.07), GCS score less than 9 (OR 19.29; 95% CI 5.04 to ∼73.82) and skull bone fracture (OR 10.44; 95% CI 3.59 to ∼30.38) were identified as possible risk factors of mortality in TBI patients.

Conclusion: These predictors appear to be clinically relevant and may help improve ED triage of TBI patients in need of ICU care.

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Head injury is a major cause of mortality. Fatality rates vary depending on the severity and mechanisms of the traumatic brain injury (TBI), but unfavourable outcomes (death, vegetative state and severe disability) following TBI can be higher than 20%. TBI has been classified (based on Glasgow coma scale (GCS) scores) into mild (GCS score 14–15), moderate (9–13), and severe (<9) TBI. Mortality is less than 20% in those with moderate TBI and approximately 40% in those with severe TBI.1 Previously identified mortality risk factors in TBI patients include hypotension, age and uncontrolled increase of intracranial pressure.24

Risk factor assessment in the emergency department (ED) is needed not only to make therapeutic and diagnostic decisions but also to aid in counselling patients and their relatives. At present, doctors lack confidence that their prognoses are meaningful. A survey of ED doctors found that only 37% of physicians thought the prognoses given to their TBI patients were accurate, suggesting that clinically useful assessment and outcome tools are critically needed.5 Although they are the first to treat trauma patients, emergency physicians have limited information regarding factors that affect the prognosis of TBI.

This retrospective study attempted to identify risk factors affecting mortality in TBI patients.


General design

To assess risk factors potentially related to mortality due to TBI in patients admitted to the ED in need of intensive care, we retrospectively reviewed the medical records of such patients who visited our ED from June 2004 to May 2005.

Patient selection

Our hospital is a tertiary care facility with a neurosurgeon and operating room available on a 24-h basis. Our ED serves approximately 70 000 patients annually. The criteria for inclusion were brain computed tomography (CT) scan and diagnosis of TBI in the ED and the need for intensive care. Patients with multiple major traumas, with penetrating brain injury, or who died on arrival were excluded. A total of 204 patients were included and were divided into two groups: the mortality group (48 patients who died in the intensive care unit (ICU) of our hospital) and the control group (156 patients who were admitted to the ICU and subsequently discharged). Initial data collected on arrival in the ED (including initial blood pressure, GCS score, systolic blood pressure, diastolic blood pressure, heart rate, etc) were analysed.

Data analysis

Age, sex, initial GCS score, initial heart rate, initial blood pressure, cause of trauma, findings of CT scan performed on ED admission, source of referral, patient transfer status and surgical intervention were chosen as predictor variables for analyses. Hypertension was defined as systolic blood pressure higher than 180 mm Hg and hypotension as systolic blood pressure lower than 90 mm Hg. All CT scan findings were taken from official radiologist reports.

Differences between the two groups were assessed using the χ2 test for categorical variables. The Mann–Whitney U test was used to test the difference for continuous variables, because the data were far from the normal distribution. The strength of risk (odds ratio) for risk factors influencing death due to TBI in the hospital was estimated by logistic regression using the enter procedure.

All statistical assessments are two-sided and evaluated at the 0.05 level of significant difference. Statistical analyses are performed using SPSS 14.0 statistics software.


There were no significant between-group differences in gender, blood pressure, heart rate, number of patients transferred from other hospitals and surgical treatment (all p>0.05). However, significantly older age (59 versus 48 years, p = 0.02), higher percentage of patients with hypotension (14.6% versus 4.5%, p = 0.02) and patients with hypertension (20.8% versus 14.1%, p<0.01) were found in the mortality group. Besides, when the initial GCS score (used as a clinical measure of the seriousness of trauma) was entered as a categorical variable, approximately 70.9% of patients in the mortality group had a GCS score lower than 9, whereas approximately 76.9% of patients in the control group had a GCS score of 9 or above (p<0.01), suggesting again that a lower score was associated with mortality (table 1).

Table 1 Clinical characteristics§

The rate of subarachnoid haemorrhage (SAH), subdural haemorrhage (SDH), brain herniation and skull bone fracture was significantly higher (all p<0.05) in the mortality group and therefore associated with mortality (72.9% versus 48.7%, p = 0.01, 77.1% versus 45.5%, p = 0.01, 8.3% versus 1.3%, p = 0.03, 70.8% versus 23.7%, p<0.01, respectively) (table 2).

Table 2 Summary of trauma aetiology and CT scan findings†‡

After controlling for the factors significantly related to outcome (all p<0.05) (tables 1 and 2), death due to TBI was more likely in patients who were older (OR 1.04, 95% CI 1.01 to 1.07, p<0.01) or who had skull fractures (OR 10.44, 95% CI 3.59 to 30.38, p<0.01). When the initial GCS score was entered as a categorical variable, the odds ratios of death for patients with a GCS score less than 9 compared with those with a GCS score above 14 were 19.29 (OR 19.29, 95% CI 5.04 to 73.82, p<0.01) (table 3).

Table 3 Multiple logistic regression analysis for factors associated with death due to TBI†


Many patients are admitted to the ED because of head injury. Patients involved in accidents frequently suffer multiple injuries. To reduce the ambiguity inherent in identifying risk factors for severe TBI in patients who need ICU treatment, we excluded patients with multiple severe traumas.

In our study, the mortality group was significantly older than the control group, suggesting that age is a risk factor for mortality in TBI patients. It has been reported that mortality as a result of TBI begins to increase at the age of 30 years.6 Moreover, mortality is higher and functional outcome poorer in older than younger patients who present with TBI, even though their head injuries and overall injuries are seemingly less severe.7

In our study, most patients in the mortality group had a GCS score lower than 9, in contrast to the control group. Similarly, the mortality rate was less than 20% compared with 40% in TBI patients with intermediate versus low GCS scores in a previous study.1 Studies have suggested that the GCS score is a risk factor for mortality in both adult (>14 years) TBI patients8 and paediatric TBI patients.2 As the GCS score changes after management (ie, after initial treatment) in the ED, the initial GCS score should be much more reliable than that after treatment.9 10 In this study, we showed that the initial GCS score could be used to predict prognosis. This would be particularly helpful when ED are backed up and their ability to provide initial treatment for conditions requiring immediate attention is affected.

Manley et al3 found that hypotension could be associated with increased mortality following brain injury. In contrast, Shafi and Gentilello11 found simply an association of hypotension with increased mortality but no association of hypotension with increased mortality due to TBI. In our study, the percentage of patients with hypotension (mortality versus control 7 (14.6%) versus 7 (4.5%)) was significantly different between groups (table 1), but no significant difference was identified by multivariate logistic regression (table 3). The relationship between mortality and hypertension was also investigated. Estimates of cerebral perfusion pressure are defined as the difference between mean arterial pressure and mean intracranial pressure. A prolonged increase in intracranial pressure is a risk factor for poor outcome in TBI patients, and a rapid rise in intracranial pressure may trigger Cushing reflex (hypertension, bradycardia, and respiratory irregularity).4 However, severe hypertension was not a risk factor for mortality in our study.

Interhospital transfer could be a risk factor for two reasons: (1) the head trauma of patients transferred from other hospitals may be more severe or (2) the condition of patients may deteriorate during transfer and be worse on arrival. One study suggested that out-of-hospital intubation among air-transported patients resulted in better outcomes than ED intubation among ground-transported patients.12 The issue is thus more complex than the mere fact of transfer alone. More research is needed to determine whether transfer is a true risk factor.

All the CT scans were performed in the ED and interpreted by the attending radiologist. Even when more than one CT scan was performed in the ED, only the first was considered. Moreover, the scan might have shown none or more than one finding. In our study, risk factors for mortality included skull bone fracture, SAH, brain herniation and SDH as seen on CT, but only skull fracture was identified by multivariate logistic regression (table 3). Thirty-four (70.8%) patients in the mortality group and 37 (23.7%) in the control group had skull fractures (p<0.01). The difference might be caused by a between-group difference in the severity of injury, which is likely to be greater in the mortality group.

In our study, skull bone fracture is a risk factor for mortality. Servadei et al13 found that skull bone fracture is a risk factor for intracranial abnormalities and the need for surgical treatment. However, few studies indicated a relationship between skull bone fracture and mortality. Mortality in patients with skull bone fracture may be higher because the fracture is caused by a more severe impact to the head.

Brain herniation was expected to be a mortality risk factor in our study, because it was a late sign of severe increased intracranial pressure. However, the absence of a significant difference by logistic regression can be attributed to the small sample size (four patients in the mortality group and two in the control group). Most of the limitations of this study stem from the relatively small sample size and its retrospective nature. The data (vital signs, GCS score, etc) might not reflect the true extent and severity of brain injury, given that these data were gathered in emergency settings. The exclusion of patients with multiple severe traumas might have contributed to selection bias, because their deaths might have been caused by trauma other than brain injury. Patients with penetrating brain injuries would have been excluded, but there were none to exclude, possibly because private gun ownership is illegal in Taiwan.

In conclusion, we found that age, skull bone fracture, and initial GCS score were risk factors for mortality in ICU-treated patients with TBI. In support of our findings, a recent study whose purpose was to formulate a decision rule for interhospital transfer of TBI patients found that five variables (initial ED GCS score, initial ED respiratory rate, initial ED temperature, initial ED pulse and patient age) could be used to identify patients requiring timely high therapeutic intensity measures and that the decision tree based on these variables was 94% sensitive and 63% specific.14 Initial blood pressure, brain herniation, SDH, SAH on initial CT scan and transfer from other hospitals may also be mortality risks, but more data are needed to confirm this possibility. Although confirmation is lacking, emergency physicians should for the present remain alert to these risk factors when managing TBI patients.



  • Competing interests: None declared.

  • Ethics approval: Ethics approval was obtained.

  • Patient consent: Obtained.

Linked Articles

  • Primary survey
    Kevin Mackway-Jones